How to delete all the lines which have blank fields with sed or awk?

Question

I would like to delete all the lines which have blank fields from text files. How can I change the following code to make the changes directly to the files?

awk '!/^\t|\t\t|\t$/' *.txt    

AD  125.9    MN 124.9
AC  38.9     VG 13.2
AV  34.6     BG 33.0
GL  126.2   
CY  34.9    
CY  44.9

desired output

AD  125.9    MN 124.9
AC  38.9     VG 13.2
AV  34.6     BG 33.0

are you mainly asking how to change the file in-place? check here: http://stackoverflow.com/questions/3979548/in-place-processing-with-grep — benroth, Oct 01 '12 at 07:17
What is a 'blank field' in your context? Please provide an example of your input and of the desired output. — , Oct 01 '12 at 07:29

score 7 · Answer 1 · answered Oct 01 '12 at 08:22

7

alternative awk one-liner:

awk 'NF==4' yourFile

answered Oct 01 '12 at 08:22

Kent

189,393
32
233
301

I don't have `awk`, but `mawk.exe` is `mawk "NF==5" myfile` possible on Windows if I would have 5 columns? – Jeroen Steen Oct 09 '19 at 19:19

score 3 · Accepted Answer · answered Oct 01 '12 at 09:45

3

One way using GNU sed:

sed -n -s -i '/[^ \t]* [^ \t]* [^ \t]* [^ \t]/p' *.txt

Results:

AD  125.9    MN 124.9
AC  38.9     VG 13.2
AV  34.6     BG 33.0

answered Oct 01 '12 at 09:45

Steve

51,466
13
89
103

Thank you very much for your answer. Your code worked correctly with the given example. But when it uses with my original files, it deletes all the lines from the files.Is there any other option for this? – user1606106 Oct 03 '12 at 13:24
@user1606106: This code suppresses printing with the `-n` flag. It considers files separately rather than as a continuous stream with the `-s` flag and edits files in place with the `-i` flag. When it finds a line with four non-empty fields or more, it will print them. So if your original files do not have four or more fields, they will be emptied. If you are able to upload some of your original files and some expected output for each to something like [dropbox](https://www.dropbox.com/) or [mediafire](http://www.mediafire.com/), perhaps I can help you. – Steve Oct 03 '12 at 22:57

score 0 · Answer 3 · answered Oct 01 '12 at 08:02

0

You could use a for loop with your original awk command:

for f in *.txt
do
  cp $f /tmp/mytempfile
  awk '!/^\t|\t\t|\t$/' /tmp/mytempfile > $f
done

answered Oct 01 '12 at 08:02

beny23

34,390
5
82
85

score 0 · Answer 4 · answered Oct 01 '12 at 08:02

0

Using your input this might work (but it depends if there are spaces or tabs):

sed '/^.\{13\}\s/d' INPUTFILE

Or this:

awk 'substr($0,14,1) = " " { print }' INPUTFILE

answered Oct 01 '12 at 08:02

Zsolt Botykai

50,406
14
85
110

I tried all the answers and none of them worked. I need to make the changes directly to the files instead of printing the output. – user1606106 Oct 01 '12 at 08:41
@user1606106 Add the `-i` flag to the `sed` option. `awk` cannot do that afaik. – Bernhard Oct 01 '12 at 09:23

Vijay · Answer 5 · 2012-10-01T08:29:43.893

0

assuming your file has a max of 4 fields:

awk '{for(i=1;i<5;i++){if($i=="")next;}print}' file_name

tested below:

> cat temp
AD      125.9    MN     124.9
AC      38.9     VG     13.2
AV      34.6     BG     33.0
GL      126.2
CY      34.9
CY      44.9
>awk '{for(i=1;i<5;i++){if($i=="")next;}print}' temp
AD      125.9    MN     124.9
AC      38.9     VG     13.2
AV      34.6     BG     33.0
>

edited Oct 01 '12 at 08:29

answered Oct 01 '12 at 08:24

Vijay

65,327
90
227
319

Thank you very much for your answer. I don't need to print the output. I would like to make the changes directly to the original files. – user1606106 Oct 01 '12 at 08:46

How to delete all the lines which have blank fields with sed or awk?

5 Answers5